OBJECTIVE: The primary objective of this project is to determine if supplementation with multiple vitamin-minerals reduces the incidence and mortality of esophageal and/or other cancers, and total mortality. Secondary objectives are to assess supplement effects on intermediate endpoints in carcinogenesis and to examine new hypotheses relating to the etiology and prevention of upper gastrointestinal cancers. BACKGROUND: Probably the highest worldwide rates of esophageal cancer occur in Linxian, China where cumulative death rates to age 75 for this cancer exceed 20%. Widespread deficiencies of multiple nutrients are considered the most likely cause, but consumption of pickled vegetables, moldy foods and nitrosamines, and physical trauma from silica fragments in ingested millet may play a role. METHODS: Two double-blind, placebo-controlled randomized intervention studies were conducted to evaluate multiple vitamins/minerals in the prevention of esophageal and other cancers. The Dysplasia Trial evaluated 3,318 persons with cytologic evidence of dysplasia supplemented for 6 years, while the General Population Trial evaluated 29,584 persons supplemented for 5 1/4 years. Biologic specimens (blood, toenails, cytology and histology specimens) were collected periodically throughout the trials. PROGRESS: Both trials concluded in 1991, and results showed that the combination of beta-carotene/vitamin E/selenium significantly reduced total mortality, total cancer mortality, and stomach cancer incidence and mortality. Results from endoscopic and cytologic examinations suggested that multiple vitamins/minerals may decrease proliferation and enhance cytologic reversion to nondysplasia. A 5-year post-intervention followup to re-examine trial results was conducted in 1996 and data analysis are in progress. A number of nested case-cohort studies relating baseline serum values of micronutrients, fumonisins, Helicobacter pylori, HPV, and EBV to cancer have recently been completed. Results of recent analyses have shown: (1) an especially strong association with increased risk for esophageal and gastric cardia cancer among subjects with low serum selenium levels; (2) increased risk for H pylori exposure for both body of stomach as well as gastric cardia cancers; and (3) no relation of fumonisin exposure with esophageal cancer risk. Gene and gene-environment interactions are currently under investigation.